The Medicaid Match That Pays States More for Adults Than for Disabled Kids

Here is a puzzle for a civics class. A state budget director has one more dollar to spend on Medicaid. She can spend it on a working-age adult with no children and no disability, or on a person with a severe disability who needs help to stay at home instead of in an institution. The federal government will add money to her dollar either way. Which choice brings in more federal money?

If you guessed the person with the disability, you have the right moral instinct and the wrong answer. Under the Affordable Care Act, the federal government pays a larger share for the adult. That is not a rumor. It is a formula, and formulas can be checked.

This post works through the arithmetic, one step at a time, because I think the numbers make the point better than adjectives. Along the way I will give the other side its best argument, since there is one, and answer it.

Terms: what a match is

Medicaid is a joint federal-state program created in 1965 as Title XIX of the Social Security Act. It is not a federal program with state offices; it is a state program with federal money. The federal government reimburses each state for a percentage of what the state spends on covered services for eligible people. That percentage is the Federal Medical Assistance Percentage, or FMAP. It is a matching grant with no ceiling. The more the state spends, the more Washington reimburses, at the same rate, with no annual cap. That structure is the important feature, and the rest follows from it.

The regular FMAP is set by a formula in the statute, and it is meant to give poorer states more help. The state's share is based on its per capita income compared with the national average:

FMAP = 100 percent minus [0.45 times (state per capita income divided by U.S. per capita income) squared]

There is a floor: no state gets less than 50 percent. There is a statutory ceiling of 83 percent, though in practice no state has been near it. In practice the formula produces a range from 50 percent up to roughly 77 percent for the poorest states in recent years, which you can verify in the FMAP table that the Department of Health and Human Services publishes in the Federal Register each year.

Let me show the formula working, using hypothetical states so that nobody is accused of anything.

• State A has per capita income equal to the national average. The ratio is 1, and 1 squared is 1. So 0.45 times 1 is 0.45, and the FMAP is 100 minus 45, which is 55 percent.
• State B has per capita income at 80 percent of the national average. The ratio is 0.8, and 0.8 squared is 0.64. So 0.45 times 0.64 is 0.288, and the FMAP is 100 minus 28.8, which is 71.2 percent.
• State C has per capita income at 120 percent of the national average. The ratio is 1.2, and 1.2 squared is 1.44. So 0.45 times 1.44 is 0.648, and the formula gives 100 minus 64.8, or 35.2 percent. That is below the floor, so the state gets 50 percent.

That is the regular match. It is the one that pays for what Medicaid was built to cover: poor children, pregnant women, parents with very low incomes, the elderly poor and people with disabilities.

The richer match

The Affordable Care Act of 2010 added a new group to the program: adults under 65 with incomes up to 133 percent of the federal poverty level, which in practice is 138 percent because of a 5 percent income disregard. Before the law, most states did not cover adults without dependent children at all, no matter how poor, unless they were disabled. The expansion changed that.

Congress gave the expansion group its own match, far above the regular one. The statute set it at 100 percent for the years 2014 through 2016, then phased down: 95 percent in 2017, 94 percent in 2018, 93 percent in 2019, and 90 percent in 2020 and every year after. The point was to make saying yes irresistible.

Now put the two matches next to each other.

Take a state at the middle of the income range, say a 60 percent FMAP (a hypothetical figure, between State A and State B above). For every $10,000 spent on a person in a regular Medicaid category, the federal government pays $6,000 and the state pays $4,000. For every $10,000 spent on an adult in the expansion group, the federal government pays $9,000 and the state pays $1,000.

The state's cost of covering the second person is one quarter of its cost for covering the first.

Another way to see it is to ask what one state dollar draws in from Washington. The federal dollars per state dollar equal FMAP divided by (1 minus FMAP).

• At a regular 60 percent FMAP: 0.60 divided by 0.40 is 1.5. Each state dollar draws $1.50.
• At the 90 percent expansion rate: 0.90 divided by 0.10 is 9. Each state dollar draws $9.
• Even for a poor state with the regular rate at 77 percent: 0.77 divided by 0.23 is about 3.35, against the same 9 for expansion adults.
• For a wealthy state at the 50 percent floor: 0.50 divided by 0.50 is 1, against 9.

Nine to one. In a rich state, expansion adults draw nine times as much federal money per state dollar as a regular enrollee does. In the poorest state, they still draw more than twice as much. The match was designed to reward a state for covering exactly one group, and that group is not the one Medicaid was first built to serve.

What that does to priorities

I am not claiming that any budget director sits down with this table and decides to starve the disabled. I am claiming that incentives work on everyone, including budget directors, and that this incentive points the wrong way.

Consider how Medicaid pays for long-term care. Nursing facility care is a mandatory benefit that states must cover. But home and community-based services, the personal care, supported living and in-home help that let a person with a disability live in an ordinary house instead of an institution, are largely optional. States provide them through waiver programs, usually under section 1915(c) of the Social Security Act, which Congress created in the early 1980s, and states can cap the number of people a waiver serves. When the cap is reached, a person goes on a waiting list.

Those waiting lists are real. The Kaiser Family Foundation has tracked them in annual surveys of the states, and they show that in a number of states, people with developmental and other disabilities wait years for the services that allow them to stay at home. I will not quote a count from memory, because the totals move from year to year; read the latest survey for yourself. What matters is the structure: the benefit for the adult in the expansion group is mandatory-in-effect, with no cap and no waiting list, while the benefit for the disabled child is rationed by slot.

The Supreme Court made the moral stakes of this plain in Olmstead v. L.C. (1999). The Court held that unjustified institutionalization of people with disabilities is a form of discrimination under the Americans with Disabilities Act, and that states must provide services in community settings when appropriate and when it can be reasonably accommodated. Olmstead says the disabled person's claim to a home is a matter of right. Yet the financing structure says the state gets nine federal dollars per state dollar for one group, and roughly one and a half for another.

Congress did partly notice. The Affordable Care Act itself included an option, the Community First Choice program, that raises the federal match for certain home-based attendant services by six percentage points. A six-point bump on a 60 percent match brings it to 66 percent, which is 0.66 divided by 0.34, or about 1.9 federal dollars per state dollar. That is progress, and it is also still a long way from nine.

What NFIB v. Sebelius did and did not do

You may ask: wasn't expansion optional? Yes, and that is the Supreme Court's doing.

The original ACA said that states that refused to expand would lose all of their existing Medicaid funding. In NFIB v. Sebelius (2012), seven Justices concluded that this was unconstitutionally coercive. Chief Justice Roberts, writing for the controlling opinion, held that the threat to take away existing Medicaid money left the states no real option but to go along. The Spending Clause lets Congress attach conditions to money it offers to states, but it does not let Congress force the states to accept conditions they could not have anticipated when they joined the program. The Court's remedy was narrow: the expansion remained in the law, but a state that declined it could not lose its existing funds.

That is a federalism victory, and I take it as one. But it changed the question without answering it. The states were free to say no, and then the question was whether the price of saying no was too high. The richer match makes it so. A state that declines expansion is turning down nine federal dollars per state dollar, while still paying federal taxes that fund the program in other states. Federalism formally survives. The incentive does the pressing.

For the broader federalism case, and what the Constitution left to the states, see Federalism Lets Fifty States Experiment, and the Constitution Planned It.

The best case for the match

The advocates for expansion, I want to be fair, have a case that deserves a hearing. They would say it this way.

The poor adult with no children was the most obviously uninsured group in the system. Medicaid had a gap that left a single adult earning a modest wage with no access to coverage, and the ACA closed it. A higher federal match was the price of getting states to participate, and that price was cheap compared with the cost of uncompensated hospital care. Many states already spent their own money on care for the uninsured through hospitals, clinics and emergency rooms, so the expansion shifted costs from state and local budgets to the federal one. And the ten percent state share is not trivial; the states do pay.

Some of that is true, and I will concede the last point. The state share is not zero. The claim I am answering is the claim that expansion is "free money for a state." That phrase appears in debates often enough to deserve a direct reply.

Is it free money? Four answers

1. It is not free to the taxpayer. The federal dollar is not a different kind of dollar. The same people pay federal taxes whether the money is spent in their state or another. If a state turns down the match, it still pays the tax that funds other states' matches. If it accepts the match, it is spending money that was borrowed when taxes fell short. The public debt was $40.56 trillion on November 5, 2026, up from $38.05 trillion a year earlier, per Treasury, an increase of $2.51 trillion. A program that draws on that is not free in any ordinary sense. See Seven myths about the national debt that Washington finds convenient.
2. The state share is ten percent of a large number. Ten percent sounds small until the base is large. A hypothetical state that enrolls 500,000 expansion adults at an average of $7,000 each spends $3.5 billion; its share at 10 percent is $350 million per year. That is not an argument against covering them. It is an argument against calling it free.
3. The 100 percent years were temporary. The federal match fell from 100 percent to 90 percent over six years. A deal that begins as free and ends at ninety cents on the dollar is one in which the state's future share rises as the federal sweetener fades, and nothing in the statute prevents a later Congress from changing it. Congress can always change the match, and states that built budgets on today's rate would have little recourse.
4. The priority problem is separate from the cost problem. Even if you believe expansion saves a state money, the question remains why the match should be richer for adults who can work than for people who cannot. If the goal is to help the most vulnerable first, the formula should say so. At a minimum, the match for home and community-based services for people with disabilities should not be lower than the match for the expansion group.

Where this leaves a conservative reformer

The right reform is not to take health coverage away from anyone. It is to fix the incentive. There are several ways, and they differ.

• Equalize the match. Pay the same FMAP for every Medicaid enrollee in a given state. That removes the thumb on the scale.
• Raise the match for the disabled and elderly in home settings. If Congress wants to buy something with a federal premium, buy the thing the program was built for.
• Block-grant or per-capita cap. Give states a fixed amount per enrollee, or per category of enrollee, and let them manage. The objection is that a cap shifts the risk of rising costs to the states, which is a fair objection and one for the states to weigh. The argument for it is that open-ended matching rewards spending and a cap rewards efficiency, the same logic that applies to Medicare, as discussed in Medicare's Two Trust Funds and the 45 Percent Warning Congress Keeps Ignoring.
• Pair coverage with work and responsibility. If the able-bodied adult group is to be covered, a work requirement for those who can work is a reasonable condition, as argued in Asking Able-Bodied Adults to Work for Medicaid Is Not Cruelty, It Is Dignity.

You can disagree with any of those. What you cannot do, if you are being honest, is say that the formula is neutral. The formula pays nine to one for one group, one and a half to one for another, and a promise that the budget director will do the morally right thing despite the arithmetic is a thin reed.

The same logic of how benefits and phase-outs shape behavior appears in the marriage penalty buried in welfare programs, with the math shown. When a program tells people what it will pay for, people respond.

The case in brief

1. Medicaid is a matching grant with no cap; the regular FMAP runs from 50 percent to about 77 percent depending on state income, by a formula in the statute.
2. The ACA paid 100 percent for the expansion group in 2014 through 2016, then phased down to 90 percent from 2020, so a state dollar draws nine federal dollars for an expansion adult against about one and a half in a mid-income state's regular match.
3. Home and community-based services for disabled people are mostly optional, capped by waiver slots, and subject to waiting lists, even though Olmstead v. L.C. (1999) treats unjustified institutionalization as discrimination.
4. NFIB v. Sebelius (2012) made expansion optional, but the richer match still presses states, and "free money" is wrong: the money is borrowed, the state share is real, and the sweetener fades.

What you can do: look up your state's FMAP in the table HHS publishes in the Federal Register, then ask your state legislators whether your state keeps a waiting list for home and community-based services and what it would cost to clear it.